Feature: Support discovery monitoring instance through http_sd

[Calvin979]

What's changed?

Corresponding issue:
#1625

Service discovery can be available for all collectors now and I provide http_sd as a deafult strategy.
There's something different between one time task and cyclic task while fetching instance info:

1. one time task
   
2. cyclic task
   

How to add service discovery for collector?

1. implements CommonProtocol interface
   
2. add param field like http_sd in yml
   

  - field: http_sd
    name:
      zh-CN: http_sd
      en-US: http_sd
    type: text
    placeholder: Enter http_sd url
    required: false
    hide: true

3. override method getSupportProtocolClass and point out which protocol is used for current collector
   

Test result

What's more
the priority of sd is higher than host, which means it will use sd first if you config host and http_sd simultaneously.

Checklist

• I have read the Contributing Guide
• I have written the necessary doc or comment.
• I have added the necessary unit tests and all cases have passed.

Add or update API

• I have added the necessary e2e tests and all cases have passed.

[Calvin979]

I haven't add sd for any collectors, it needs to be discussed which collector should be supported for sd

[tomsun28]

👍👍 This design implementation is eye-catching and is a good idea. I don't seem to find the part that automatically discovers and creates multiple instances through the information returned by the http_sd interface. For example, http_sd returns ip1-port1 ip2-port2 ip3-port3. We create three new monitoring instances based on these three automatic discoveries.

[Calvin979]

@tomsun28
I'm not sure whether you mean we need to create monitor task automatically. If so, the process is different.
Monitor task must be created and managed by Manager, then collector should collect sd data as a cyclic task like this:

In this case I think batch create task would be better, it will change a lot of code and design.

What I coded in this pr is update host and port automatically.

It will be helpful for monitoring services that are supported for automatic failover like Redis sentinel.

[tomsun28]

hi yes, the first design like this as same as your.

Also i think this pr design is great, provides another way of thinking. how about combine the two. Our initial goal is to support automatic maintenance of batch monitoring tasks through the data returned by http_sd.

[Calvin979]

@tomsun28
Thanks for your design. It is a more clear design than my design. Should I modify code in this pr directly?🤣🤣

[tomsun28]

👍👍 okay okay. It is not urgent, depends on your time. Recommend to release an first apache version first and after then consider integrate this big feature.

[Calvin979]

👍👍 okay okay. It is not urgent, depends on your time. Recommend to release an first apache version first and after then consider integrate this big feature.

OK, I will close this pr. And I will open another pr after first apache version released and I finished this new feature.